Open leemartin opened 9 years ago
Hi. I am trying to do the same thing that you propose, but I have any solution yet. If anybody knows how to resolve this problem, it will be great to improve the result of the eye detection and to reduce te computational cost.
Is there anyway to use the detected face rect as an ROI for eye detection? I suspect that would speed things up on the eye detection side quite a bit. Also, any pointers on detecting which eye is present: left or right. I was going to use the face rect and decide if the eye was positioned to the left or right (within the face rect box) but curious if you know of a better solution. Thanks!